I was recently reading a thread here about a person who was getting 100 degree C temperatures from running FurMark at stock clocks (that was an isolated issue that I think is resolved now). In that thread though, I saw several people arguing that FurMark is not a useful test and only serves to damage hardware. While I have always known it is not representative of real usage, I had not heard before that it should be avoided. I personally use it for a couple minutes at a time to get an idea if my core clock and voltage are remotely stable or not, then proceed to run benchmark type tests (3DMark) for a more realistic stress test.
In your opinion:
Is FurMark useful?
Are you concerned about hardware lifespan from it?
More generally, what do you use for stability testing?